Attribution analysis method and device, equipment and storage medium

By performing unified preprocessing of log data from multiple log platforms and attribution analysis of natural language text input generative models, the problems of low efficiency and insufficient accuracy of attribution analysis in the existing technology are solved, and more efficient and stable are achieved Location of fault problems.

CN119046048APending Publication Date: 2024-11-29BEIJING ZITIAO NETWORK TECH CO LTD
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202411196095.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

In the prior art, the attribution analysis process for functional program failure problems is time-consuming, inefficient, and the accuracy and stability of the results are insufficient by relying on manual experience.

Method used

By extracting log data from multiple log platforms for unified preprocessing, serialized log text is generated, and integrated it with the problem description into natural language text, input a generative model for attribution analysis, and generating attribution results.

Benefits of technology

Improve the efficiency and accuracy of attribution analysis, reduce manual participation, and enhance the stability of the final result.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046048A_ABST
    Figure CN119046048A_ABST
Patent Text Reader

Abstract

The embodiment of the invention relates to an attribution analysis method and device, equipment and a storage medium. The method comprises the steps of determining a user identifier, problem occurrence time and problem description in response to an attribution analysis request; extracting first log data from a plurality of log platforms based on the user identifier and the problem occurrence time; preprocessing the first log data to generate a serialized log text, and generating a natural language text to be analyzed based on the serialized log text and the problem description; and inputting the natural language text to be analyzed into a generative model to generate an attribution result corresponding to the attribution analysis request. According to the embodiment of the invention, the problem description and the log data can be input into the generative model in the form of the natural language text, and the understanding ability and the attribution ability of the model to the problem description and the log data are improved, so that the accuracy and the stability of an attribution result are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to an attribution analysis method, apparatus, device and storage medium. Background Art

[0002] When a problem occurs in an online functional program, the user or program will report relevant information about the problem, such as user ID, problem description, and time when the problem occurred. Relevant personnel will extract multiple log data related to the problem from various log platforms based on the user ID. Then, relevant personnel will organize these log data, combine them with the problem description, and analyze and troubleshoot these log data one by one based on their own experience and familiarity with the functional program, so as to locate the cause of the above problem.

[0003] However, the above process requires relevant personnel to manually analyze and troubleshoot a large amount of log data, which makes the attribution analysis process of the problem time-consuming and inefficient. In addition, due to the extreme reliance on the experience of relevant personnel, the accuracy and stability of the causes located after the attribution analysis are insufficient. Summary of the invention

[0004] In order to solve the above technical problems, the embodiments of the present disclosure provide an attribution analysis method, apparatus, device and storage medium.

[0005] In a first aspect, an embodiment of the present disclosure provides an attribution analysis method, the method comprising:

[0006] In response to the attribution analysis request, determine the user identification, the time when the problem occurred, and the description of the problem;

[0007] Extracting first log data from multiple log platforms based on the user identifier and the time when the problem occurred;

[0008] Preprocessing the first log data to generate serialized log text, and generating a natural language text to be analyzed based on the serialized log text and the problem description;

[0009] The natural language text to be analyzed is input into a generative model to generate an attribution result corresponding to the attribution analysis request.

[0010] In a second aspect, an embodiment of the present disclosure further provides an attribution analysis device, the device comprising:

[0011] A problem description determination module, for determining a user identifier, a problem occurrence time, and a problem description in response to an attribution analysis request;

[0012] A first log data extraction module, used to extract first log data from multiple log platforms based on the user identifier and the problem occurrence time;

[0013] A natural language text generation module, used to preprocess the first log data to generate serialized log text, and generate a natural language text to be analyzed based on the serialized log text and the problem description;

[0014] The attribution result generation module is used to input the natural language text to be analyzed into the generative model to generate the attribution result corresponding to the attribution analysis request.

[0015] In a third aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0016] processor;

[0017] A memory for storing executable instructions;

[0018] The processor is used to read executable instructions from the memory and execute the executable instructions to implement the attribution analysis method described in any embodiment of the present disclosure.

[0019] In a fourth aspect, an embodiment of the present disclosure further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements the attribution analysis method described in any embodiment of the present disclosure.

[0020] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, which is used to execute the attribution analysis method described in any embodiment of the present disclosure.

[0021] The attribution analysis method, apparatus, device and storage medium of the disclosed embodiment can determine the user identification, problem occurrence time and problem description from the attribution analysis request; based on the user identification and the problem occurrence time, extract the first log data from multiple log platforms; pre-process the first log data to generate a serialized log text, and based on the serialized log text and the problem description, generate a natural language text to be analyzed; input the natural language text to be analyzed into a generative model to generate an attribution result corresponding to the attribution analysis request; realize the unified pre-processing of log data extracted from different log platforms , reducing the differences between log data, improving the correlation of log data, saving the time required for manual sorting of log data, and thus improving the efficiency of attribution analysis to a certain extent; in addition, the preprocessed log data is text-processed, and the obtained natural language text is input into the generative model, so that the model can give full play to its natural language processing capabilities, so as to better understand the problem description and log data in the natural language text, so as to better perform attribution analysis on the log data, avoiding the time-consuming manual attribution analysis and excessive reliance on manual experience, and improving the accuracy and stability of the final attribution results.

[0022] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present disclosure are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0024] Figure 1 A schematic diagram of a process of an attribution analysis method provided by an embodiment of the present disclosure;

[0025] Figure 2 A flowchart of an attribution analysis method provided by an embodiment of the present disclosure;

[0026] Figure 3 for Figure 2 A detailed flow chart of S230 in the attribution analysis method shown;

[0027] Figure 4 A schematic diagram of another attribution analysis method provided by an embodiment of the present disclosure;

[0028] Figure 5 A schematic diagram of the structure of an attribution analysis device provided by an embodiment of the present disclosure;

[0029] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0030] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0031] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0032] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0033] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0034] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0035] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0036] In the related art, when a user uses a functional program (such as an application app or a web page, etc.) or a relevant person tests a functional program, a functional failure may occur. In this case, the user or relevant person can consult customer service to inquire about the solution to the problem, or report the problem. After the maintenance personnel / developers of the functional program receive the problem and related information, they can analyze the cause of the problem by checking multiple log data related to the problem.

[0037] See also Figure 1 In the related art, after receiving the attribution analysis request, the maintenance personnel / developers can extract multiple log data from multiple log platforms according to the user ID and the time when the problem occurred. Then, relying on personal experience and business familiarity with the functional program, combined with the problem description, these log data are manually analyzed and checked one by one to finally determine the cause of the fault problem. However, this process requires a lot of manual operations and is extremely dependent on manual experience, which makes the attribution analysis process time-consuming and inefficient, and the accuracy and stability of the attribution results are poor.

[0038] Based on the above situation, continue to see Figure 1The disclosed embodiment provides an attribution analysis solution, which automatically processes and serializes multiple log data extracted from multiple log platforms, and organizes the problem description and the serialized log data to obtain a text for attribution analysis of the fault problem to be performed described in natural language (i.e., the natural language text to be analyzed). Afterwards, the natural language text to be analyzed is input into a generative model with natural language processing capabilities for attribution analysis to obtain an attribution result. In this way, since the degree of manual participation is greatly reduced, the accuracy and stability of the attribution results can be improved, and the efficiency of the attribution analysis is greatly improved.

[0039] The attribution analysis method provided by the embodiments of the present disclosure can be applied to the scenario of attribution analysis of fault problems of functional programs. For example, it can be applied to the scenario of attribution analysis of fault problems actively reported by users or relevant personnel; for another example, it can be applied to the scenario of attribution analysis of the consulted fault problems in the scenario where users and customer service consult about fault problems through dialogue; for another example, it can be applied to the scenario of attribution analysis of fault alarms generated by the alarm mechanism corresponding to the functional program. The method can be executed by an attribution analysis device, which can be implemented by software and / or hardware, and the device can be integrated in an electronic device with certain computing power. The electronic device can include, but is not limited to, smart phones, tablet computers (Tablet Personal Computer, Tablet PC), laptop computers, desktop computers, servers, server clusters, etc.

[0040] Figure 2 FIG. 1 is a flow chart of an attribution analysis method provided by an embodiment of the present disclosure. Figure 2 As shown, the attribution analysis method may include the following steps:

[0041] S210: In response to the attribution analysis request, determine the user ID, the time when the problem occurred, and the problem description.

[0042] Among them, the attribution analysis request is a request for triggering the execution of the processing flow for problem location and problem attribution of the fault problem. It can be initiated by the user of the online functional program, or by the relevant personnel of the functional program in the test environment, or by the user of the attribution analysis platform based on the data received. The user ID is the unique ID of the user who used the functional program when the fault problem occurred. The problem occurrence time is the moment when the functional program fault problem occurred. The problem description is a text expressing the fault problem in natural language.

[0043] Specifically, as described above, when using a functional program, the user can consult online customer service about the functional problems they encounter. During this process, the client corresponding to the functional program can package information such as the user ID, the description of the problem consulted by the user, the time when the functional problem occurred, the screenshot or screen recording video when the functional problem occurred, etc., construct an attribution analysis request, and send it to the electronic device used for attribution analysis. Alternatively, when a user encounters a functional problem while using a functional program, the user manually or the client automatically triggers an error reporting process to send the above information to the client's back-end server. The back-end personnel can initiate an attribution analysis request to the electronic device based on the information received.

[0044] After receiving the attribution analysis request, the electronic device can parse it to obtain the user ID, problem occurrence time and problem description contained therein.

[0045] S220: Extract first log data from multiple log platforms based on the user identifier and the time when the problem occurs.

[0046] The first log data is original log data extracted from the log platform.

[0047] Specifically, the time when the problem occurs is the moment when the functional problem occurs, but the user operation or the use of the functional program is a coherent process with contextual information. Therefore, in order to more accurately locate the event where the problem occurs, the electronic device can expand the period for obtaining log data to obtain log data before and after the functional problem occurs. For example, taking the time when the problem occurs as the central moment, both forward and backward are expanded for a certain length of time (such as 10 minutes) to determine the time period for log extraction. Then, the electronic device extracts multiple log data from multiple log platforms corresponding to the functional program according to the user identifier as the first log data.

[0048] These log platforms can be any platform that has data access rights and contains log data related to the functional program. For example, it can be a platform that collects log data corresponding to the business tracking points of the functional program, or a platform that collects log data of the technical tracking points of the web page of the functional program, or a platform that collects log data of the technical tracking points of the client of the functional program, etc.

[0049] S230: Preprocess the first log data to generate serialized log text, and generate a natural language text to be analyzed based on the serialized log text and the problem description.

[0050] The serialized log text is the result of serializing the log data in a text form, for example, the result of processing the log data using a text-based serialization algorithm.

[0051] Specifically, in the related art, attribution is performed by manually analyzing the first log data. However, the first log data is large and chaotic, such as inconsistent data formats, events are not connected in series, events may not be related to the fault problem, etc., which makes log analysis difficult. Based on this, in the embodiment of the present disclosure, the first log data can be preprocessed first to eliminate the above problems to a large extent. For example, the first log data can be preprocessed by event concatenation, data format unification, log data filtering, and text-based serialization to obtain a serialized log text.

[0052] Then, considering that the generative model based on language processing is better at processing natural language text, the embodiment of the present disclosure integrates the serialized log text and the problem description into a natural language text with relatively complete sentences and semantics as the natural language text to be analyzed. The natural language text to be analyzed will be used as the input data of the generative model, which is more compatible with the model capabilities and can improve the model's ability to understand and analyze the problem description and log data, thereby improving the accuracy of the model in attribution analysis.

[0053] S240: Input the natural language text to be analyzed into the generative model to generate an attribution result corresponding to the attribution analysis request.

[0054] Among them, the attribution result is the result obtained after locating and attributing the fault problem in the problem description.

[0055] Specifically, the electronic device uses the natural language text to be analyzed as input data, calls the generative model, and uses the natural language processing capability of the generative model to perform log analysis, attribute the problem description, and obtain the attribution result.

[0056] In some embodiments, the attribution result includes at least one of target log data related to the problem description, a cause of the problem corresponding to the problem description, and problem solution information for the problem description.

[0057] The above-mentioned target log data is the log data most relevant to the problem description, which can be the preprocessed log data or the first log data. The cause of the above-mentioned problem is the cause of the problem description generated by the generative model based on the target log data, its context log data and related context information (such as related information base / knowledge base), which is in the form of natural language text so that relevant personnel can accurately understand the target log data and the cause of the fault problem. The above-mentioned problem-solving information is advisory information on how to eliminate the fault problem, such as suggestions for improvement in product design, or suggestions for improvement in front-end / back-end development technology, etc.

[0058] Specifically, in order to enable relevant personnel to have a deeper understanding of the fault problem and its attribution, in this embodiment, the attribution result output by the generative model can be designed to include at least one of the target log data, the cause of the problem, and the problem solving information. In this way, relevant personnel can deeply understand the fault problem and locate the cause from the perspective of the log and / or cause dimension. In addition, through the problem solving information, it is possible to learn how to solve the problem more quickly, which enhances the flexibility and efficiency of problem solving.

[0059] For example, when the attribution results include target log data, problem causes, and problem resolution information, the output of the generative model can be as follows:

[0060] Target log data: [2024-07-01 20:58:38,xxx_confirm_error_info,xxx interface call,INFO,"xxx"] / / [specific date and time, "event identifier", "event description",INFO,"specific content of log data"];

[0061] Cause of the problem: Due to xxxx reasons, xxx problem occurred;

[0062] Problem solving information:

[0063] 1. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx;

[0064] 2. xxxxxxxxxxxxxxxxxxxxx;

[0065] 3. xxxxxxxxxxxxxxxxxxxx.

[0066] In some embodiments, S240 includes: integrating the natural language text to be analyzed with preset prompt words to generate target prompt words; running a generative model based on the target prompt words to generate attribution results.

[0067] The preset prompt is a model prompt that is designed in advance according to the attribution business and adjusted during the training process, which at least includes the receiving method of the input data, and may also include the input data format, output data format, logic / process / analysis key points / analysis examples of the attribution analysis, etc. The receiving method of the above input data can be a placeholder for the input data, or the input data can be directly spliced ​​with the preset prompt.

[0068] Specifically, before executing the above-mentioned attribution analysis method, relevant personnel can design the prompt word prompt (i.e., the initial prompt word) of the generative model in advance according to the attribution business. Then, collect the problem description sample and the corresponding first log data, and perform the above-mentioned preprocessing on the first log data to obtain the corresponding serialized log text sample, and the natural language text sample is composed of the problem description sample and the serialized log text sample. At the same time, mark the attribution result sample corresponding to the problem description sample. Afterwards, call the generative model with the natural language text sample and the initial prompt word to perform model training to obtain the attribution result output by the model. According to the difference between the attribution result output by the training and the attribution result sample, adjust the initial prompt word. Repeat the above-mentioned model training process until the preset end condition is reached, and the preset prompt word can be obtained.

[0069] In specific implementation, the electronic device integrates the natural language text to be analyzed into the preset prompt word according to the receiving method of the input data specified in the preset prompt word, and obtains the target prompt word that is finally used to call the generative model. For example, the electronic device can use the natural language text to be analyzed to replace the placeholder of the input data in the preset prompt word to obtain the target prompt word. For another example, the electronic device can splice the natural language text to be analyzed to the preset prompt word to obtain the target prompt word. Then, the electronic device can call the generative model through the target prompt word, and output the attribution result after model processing.

[0070] In some embodiments, after S240, the attribution analysis method further includes: if the attribution result is wrong, determining a correction result corresponding to the attribution result, and using the natural language text to be analyzed and the correction result as training samples to train the generative model to continue adjusting the preset prompt words.

[0071] Specifically, considering that the model training stage cannot take into account various actual situations, there may be a situation where the model cannot correctly attribute a certain fault problem. Therefore, this embodiment can use the fault problems attributed incorrectly as new training data, and continuously strengthen the preset prompt words to improve the accuracy of model attribution.

[0072] In specific implementation, after the model outputs the attribution result, it can be handed over to relevant personnel for verification. If the relevant personnel determine that a certain attribution result is wrong, the electronic device can receive the correct attribution result (i.e., correction result) of the fault problem given by the relevant personnel. Then, the natural language text to be analyzed corresponding to the attribution result and the correction result are used as new training samples, and the model training process of the above prompt word adjustment is performed to obtain a new prompt word to update the preset prompt word.

[0073] The attribution analysis method provided by the above-mentioned embodiments of the present disclosure can determine the user identifier, the time when the problem occurred, and the problem description from the attribution analysis request; based on the user identifier and the time when the problem occurred, extract the first log data from multiple log platforms; pre-process the first log data to generate a serialized log text, and based on the serialized log text and the problem description, generate a natural language text to be analyzed; input the natural language text to be analyzed into a generative model to generate an attribution result corresponding to the attribution analysis request; realize unified pre-processing of the log data extracted from different log platforms, reduce the difference between the log data, improve the relevance of the log data, save the time required for manual sorting of the log data, and thus improve the efficiency of the attribution analysis to a certain extent; in addition, the pre-processed log data is text-processed, and the obtained natural language text is input into the generative model, so that the model can give full play to its natural language processing ability, so as to better understand the problem description and log data in the natural language text, so as to better perform attribution analysis on the log data, avoid the time-consuming and excessive reliance on manual experience caused by manual attribution analysis, and improve the accuracy and stability of the attribution result finally obtained.

[0074] In some embodiments, Figure 3 Shows Figure 2 The detailed flow chart of S230 in the attribution analysis method shown in FIG. Figure 3 As shown, in S230, “preprocessing the first log data to generate serialized log text” includes:

[0075] S331. Arrange the first log data from a timestamp dimension to generate second log data.

[0076] Specifically, see Figure 4 , the electronic device accesses log platform 1, log platform 2, ..., log platform n according to the user identifier and problem occurrence time in the attribution analysis request, and obtains the first log data of multiple platforms. Because these log data have problems such as inconsistent data formats and unrelated events, the electronic device can first connect the log events in the first log data in series from the timestamp dimension, and the result of the arrangement is the second log data. The second log data contains complete log data and associates the log data of the same event to provide context log data for the event.

[0077] S332. Filter the second log data based on the attribution task information to determine the third log data.

[0078] Among them, attribution task information is relevant information of the task requirements of attribution analysis. For example, when the business party is more concerned about a certain business function, attribution analysis can be performed on the business function to effectively promote the optimization of the business function. At this time, the attribution task information can be the business information of the business function (such as business identification, component identification related to the business function, etc.). For another example, when the developer is more concerned about a certain front-end / back-end technical function, attribution analysis can be performed on the technical function to effectively promote the optimization of the technical function. At this time, the attribution task information can be relevant information of the technical function (such as component identification related to the technical function, etc.). For another example, when the business function corresponding to the web page side / application side / a certain type of user / a certain type of device is being promoted, the attribution task information can be the relevant identification corresponding to the web page side / application side / a certain type of user / a certain type of device, etc.

[0079] Specifically, the electronic device can obtain the attribution task information in advance, and then filter the second log data according to the attribution task information to remove the log data in the second log data that is not related to the attribution task information. The result is as follows: Figure 4 The third log data shown. In this way, the amount of data processing of the subsequent model can be reduced, the attribution efficiency can be improved, and the interference of irrelevant log data can be reduced, and the accuracy of attribution can be improved. In addition, because the second log data has been processed for event association according to the timestamp, it helps the electronic device to perform more accurate log filtering according to the context log data during the filtering process, avoiding the erroneous elimination of related log data.

[0080] In some embodiments, the attribution task information may be manually specified in advance by relevant personnel, for example, in the form of a configuration file.

[0081] In some other embodiments, the attribution task information may be automatically determined. The process includes: if the attribution analysis request also includes auxiliary information of the problem description, then determining the attribution task information adapted to the problem description based on the problem description and the auxiliary information.

[0082] Among them, auxiliary information is information used to assist in problem attribution, such as contextual information of the conversation between the user and customer service, or screenshots or screen recordings of the fault problem.

[0083] Specifically, the electronic device can parse the attribution analysis request and obtain the auxiliary information therein. Then, the attribution task information corresponding to the problem description is determined by analyzing the auxiliary information. For example, the electronic device can enable the relevant model with the dialogue analysis function in the relevant technology, and perform dialogue analysis on the problem description of the dialogue in the auxiliary information and its context information to determine whether it is a problem for the business function corresponding to the web page / application end / a certain type of user / a certain type of device, or whether it is a problem for the development technology of the front end / back end, etc. When the judgment result is yes, the attribution task information is determined in combination with the relevant context information (such as the relevant information library / knowledge base). For another example, the electronic device can use the relevant algorithm with the image analysis / video analysis function to analyze the image or video to identify whether this attribution analysis is a problem for the business function corresponding to the web page / application end / a certain type of user / a certain type of device. If yes, the attribution task information is determined in combination with the relevant context information (such as the relevant information library / knowledge base); if not, it can be considered that this attribution analysis is a problem for the development technology of the front end / back end, and the attribution task information is determined in combination with the relevant context information (such as the relevant information library / knowledge base). This can further improve the automation level of attribution analysis, thereby further improving the efficiency of attribution analysis.

[0084] S333: Serialize the third log data to generate a serialized log text.

[0085] Specifically, the electronic device may utilize a text-based serialization algorithm in the related art to perform serialization processing on the third log data, and generate a serialized log text in a text form with a unified format.

[0086] In some embodiments, S333 includes: based on a preset data format, format conversion processing is performed on the third log data to generate fourth log data; based on business parameters and non-business parameters, parameter classification and organization of the fourth log data is performed to generate fifth log data; based on the parameter type corresponding to the attribution task information, text-based serialization processing is performed on the fifth log data to generate serialized log text.

[0087] The preset data format is a pre-set data format of log data. Business parameters are parameters related to business functions. Non-business parameters are parameters unrelated to business functions, such as user ID, client ID, etc.

[0088] Specifically, see Figure 4After obtaining the third log data, the electronic device may first unify the data format of these log data to improve the efficiency of subsequent log data preprocessing. Therefore, the electronic device may convert the third log data into a preset data format, and remove the redundant parameter values ​​therein or merge them into similar parameter fields, and supplement the missing parameter values ​​therein with default values ​​or mark them as missing, and the result is the fourth log data. Then, the electronic device may classify the parameters in the fourth log data according to business parameters and non-business parameters to generate parameter-classified fifth log data. In this way, on the one hand, the meaning of the log data / parameters can be made clearer to further enhance the ability of the subsequent generative model to understand the log data; on the other hand, it can provide a basis for secondary log screening.

[0089] Then, the log data can be screened again in combination with the attribution task information. For example, if the attribution task information is information related to the xx business function, then the corresponding parameter type can be a business parameter, and the electronic device can remove the relevant data of non-business parameters in the fifth log data. For another example, if the attribution task information is information other than the business function, then the corresponding parameter type can be a non-business parameter, and the electronic device can remove the relevant data of business parameters in the fifth log data. For another example, if the attribution task information involves both information about the xx business function and information about non-business functions related to the business function, then the corresponding parameter type can be a business parameter and a non-business parameter, and the electronic device retains all the fifth log data. Afterwards, the electronic device can perform text-based serialization processing on the second-screened fifth log data to obtain a serialized log text. This can further streamline the log data, reduce redundant log data, and improve the standardization of log data, thereby providing more accurate input log data for subsequent attribution analysis, thereby further improving the accuracy and efficiency of the attribution analysis.

[0090] Continue to see Figure 4 ,In the subsequent process, the electronic device can integrate the serialized log text and ,problem description to obtain the natural language text to be analyzed. ,The natural language text to be analyzed and the preset prompt words are input ,into the generative model, which can output the attribution ,results including the target log data, the cause of the problem, and the problem ,solution information.

[0091] The following is an embodiment of an attribution analysis device provided in an embodiment of the present invention. The device and the attribution analysis methods of the above-mentioned embodiments belong to the same inventive concept. For details not fully described in the embodiment of the attribution analysis device, reference can be made to the embodiment of the above-mentioned attribution analysis method.

[0092] Figure 5 FIG. 2 shows a schematic diagram of the structure of an attribution analysis device provided by an embodiment of the present disclosure. Figure 5 As shown, the attribution analysis device 500 may include:

[0093] A problem description determination module 510 is used to determine a user identifier, a problem occurrence time, and a problem description in response to an attribution analysis request;

[0094] A first log data extraction module 520, for extracting first log data from multiple log platforms based on the user identifier and the time when the problem occurred;

[0095] A natural language text generation module 530 is used to preprocess the first log data to generate a serialized log text, and to generate a natural language text to be analyzed based on the serialized log text and the problem description;

[0096] The attribution result generation module 540 is used to input the natural language text to be analyzed into the generative model to generate the attribution result corresponding to the attribution analysis request.

[0097] The attribution analysis device provided by the above-mentioned embodiments of the present disclosure can determine the user identifier, the time when the problem occurred, and the problem description from the attribution analysis request; based on the user identifier and the time when the problem occurred, extract the first log data from multiple log platforms; pre-process the first log data to generate a serialized log text, and based on the serialized log text and the problem description, generate a natural language text to be analyzed; input the natural language text to be analyzed into a generative model to generate an attribution result corresponding to the attribution analysis request; realize unified pre-processing of the log data extracted from different log platforms, reduce the differences between the log data, improve the relevance of the log data, save the time required for manual sorting of the log data, and thus improve the efficiency of attribution analysis to a certain extent; in addition, the pre-processed log data is text-processed, and the obtained natural language text is input into the generative model, so that the model can give full play to its natural language processing ability, so as to better understand the problem description and log data in the natural language text, so as to better perform attribution analysis on the log data, avoid the time-consuming and excessive reliance on manual experience caused by manual attribution analysis, and improve the accuracy and stability of the attribution result finally obtained.

[0098] In some embodiments, the natural language text generation module 530 includes:

[0099] A second log data generating submodule, used to sort the first log data from a timestamp dimension to generate second log data;

[0100] A third log data generating submodule is used to filter the second log data based on the attribution task information to determine the third log data;

[0101] The serialized log text generation submodule is used to perform serialization processing on the third log data to generate serialized log text.

[0102] In some embodiments, the serialized log text generation submodule is specifically used to:

[0103] Based on a preset data format, performing format conversion processing on the third log data to generate fourth log data;

[0104] Based on the business parameters and non-business parameters, the fourth log data is classified and sorted by parameters to generate fifth log data;

[0105] Based on the parameter type corresponding to the attribution task information, text-based serialization processing is performed on the fifth log data to generate a serialized log text.

[0106] In some embodiments, the attribution result generation module 540 is specifically used to:

[0107] Integrate the natural language text to be analyzed with the preset prompt words to generate the target prompt words;

[0108] The generative model is run based on the target prompt word to generate the attribution result; the preset prompt word is determined in advance by iteratively adjusting the initial prompt word during the training process of the generative model using natural language text samples consisting of problem description samples and serialized log text samples and attribution result samples.

[0109] In some embodiments, the attribution analysis device 500 further includes a prompt word adjustment module, which is used to:

[0110] After the natural language text to be analyzed is input into the generative model and the attribution result corresponding to the attribution analysis request is generated, if the attribution result is wrong, the correction result corresponding to the attribution result is determined, and the natural language text to be analyzed and the correction result are used as training samples to train the generative model to continue adjusting the preset prompt words.

[0111] In some embodiments, the attribution result includes at least one of target log data related to the problem description, a cause of the problem corresponding to the problem description, and problem solution information for the problem description.

[0112] In some embodiments, the natural language text generation module 530 further includes an attribution task information determination submodule, which is used to:

[0113] Before filtering the second log data based on the attribution task information and determining the third log data, if the attribution analysis request also includes auxiliary information of the problem description, attribution task information adapted to the problem description is determined based on the problem description and the auxiliary information.

[0114] The attribution analysis device provided in the embodiment of the present invention can execute the attribution analysis method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0115] It is worth noting that in the embodiment of the attribution analysis device mentioned above, the various modules and sub-modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional modules / sub-modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present disclosure.

[0116] The present disclosure also provides an electronic device, which may include a processor and a memory, wherein the memory may be used to store executable instructions. The processor may be used to read the executable instructions from the memory and execute the executable instructions to implement the attribution analysis method in the above embodiment.

[0117] Figure 6 A schematic structural diagram of an electronic device provided by an embodiment of the present disclosure is shown.

[0118] like Figure 6 As shown, the electronic device 600 may include a processing device 601 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output interface (I / O interface) 605 is also connected to the bus 604.

[0119] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data.

[0120] It should be noted that Figure 6 The electronic device 600 shown is only an example and should not limit the functions and scope of use of the embodiments of the present disclosure. Figure 6 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0121] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the attribution analysis method of any embodiment of the present disclosure are executed.

[0122] An embodiment of the present disclosure further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the processor implements the attribution analysis method in any embodiment of the present disclosure.

[0123] It should be noted that the computer-readable medium disclosed above may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0124] In some embodiments, the client and server may communicate using any currently known or future developed network protocol such as HTTP, and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.

[0125] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0126] The computer-readable medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device executes the steps of the attribution analysis method described in any embodiment of the present disclosure.

[0127] In embodiments of the present disclosure, computer program code for performing the operations of the present disclosure may be written in one or more programming languages ​​or a combination thereof, including but not limited to object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on a user's computer, partially on a user's computer, as a separate software package, partially on a user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0128] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the equipment, method and computer program product according to various embodiments of the present disclosure. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0129] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0130] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0131] The above description is only a preferred embodiment of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present disclosure is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in the present disclosure (but not limited to) by each other to form a technical solution.

[0132] In addition, although each operation is described in a specific order, this should not be understood as requiring these operations to be performed in the specific order shown or in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present disclosure. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.

[0133] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.

Claims

1. An attribution analysis method, characterized in that: include: In response to the attribution analysis request, determine the user identification, the time when the problem occurred, and the description of the problem; Extracting first log data from multiple log platforms based on the user identifier and the time when the problem occurred; Preprocessing the first log data to generate serialized log text, and generating a natural language text to be analyzed based on the serialized log text and the problem description; The natural language text to be analyzed is input into a generative model to generate an attribution result corresponding to the attribution analysis request.

2. The method according to claim 1, characterized in that The preprocessing of the first log data to generate a serialized log text includes: Arrange the first log data based on the timestamp dimension to generate second log data; Based on the attribution task information, filtering the second log data to determine third log data; The third log data is serialized to generate the serialized log text.

3. The method according to claim 2, characterized in that The serializing the third log data to generate the serialized log text includes: Based on a preset data format, performing format conversion processing on the third log data to generate fourth log data; Based on the business parameters and non-business parameters, the fourth log data is classified and sorted by parameters to generate fifth log data; Based on the parameter type corresponding to the attribution task information, the fifth log data is subjected to text-based serialization processing to generate the serialized log text.

4. The method according to claim 1, characterized in that: The step of inputting the natural language text to be analyzed into a generative model to generate an attribution result corresponding to the attribution analysis request includes: Integrating the natural language text to be analyzed with preset prompt words to generate target prompt words; The generative model is run based on the target prompt word to generate the attribution result; the preset prompt word is determined in advance by iteratively adjusting the initial prompt word during the training of the generative model using natural language text samples consisting of problem description samples and serialized log text samples and attribution result samples.

5. The method according to claim 4, characterized in that After inputting the natural language text to be analyzed into the generative model to generate the attribution result corresponding to the attribution analysis request, the method further includes: If the attribution result is wrong, a correction result corresponding to the attribution result is determined, and the natural language text to be analyzed and the correction result are used as training samples to train the generative model, so as to continue to adjust the preset prompt word.

6. The method according to any one of claims 1 to 5, characterized in that: The attribution result includes at least one of target log data related to the problem description, a cause of the problem corresponding to the problem description, and problem solution information for the problem description.

7. The method according to claim 2, characterized in that Before filtering the second log data based on the attribution task information to determine the third log data, the method further includes: If the attribution analysis request further includes auxiliary information of the problem description, the attribution task information adapted to the problem description is determined based on the problem description and the auxiliary information.

8. An attribution analysis device, characterized in that: include: A problem description determination module, for determining a user identifier, a problem occurrence time, and a problem description in response to an attribution analysis request; A first log data extraction module, used to extract first log data from multiple log platforms based on the user identifier and the problem occurrence time; A natural language text generation module, used to preprocess the first log data to generate serialized log text, and generate a natural language text to be analyzed based on the serialized log text and the problem description; The attribution result generation module is used to input the natural language text to be analyzed into the generative model to generate the attribution result corresponding to the attribution analysis request.

9. An electronic device, characterized in that: include: processor; A memory for storing executable instructions; The processor is used to read the executable instructions from the memory and execute the executable instructions to implement the attribution analysis method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the processor implements the attribution analysis method described in any one of claims 1 to 7.

Citation Information

Cited By

  • Regression data analysis method based on unsupervised learning, computer equipment and medium

    CN121093306A

  • Regression data analysis method based on unsupervised learning, computer device and medium

    CN121093306B

  • Method and apparatus for managing behavior of application, device, and medium

    WO2026123265A1